Towards Interactively Improving ML Data Preparation Code via "Shadow Pipelines"

Keyphrases

• Data Scientist 100%
• Data Preparation 100%
• Manual Processes 50%
• Debugging 50%
• Potential Gains 50%
• Development Cycle 50%
• ML pipeline 50%
• Error-prone 50%
• Iterative Manner 50%
• Incremental View Maintenance 50%
• Low Latency 50%

Computer Science

• Data Scientist 100%
• Data Preparation 100%
• Manual Process 50%
• Potential Benefit 50%
• Development Cycle 50%
• Interactivity 50%